Indeed. It may be true that coding jobs on low skill level are often Java these days, but C is not vanishing at all. For example, I recently built a custom security filter component for a customer and while in theory it is possible to do this in Java (I think), it would be a lot of additional effort, the component would not perform well and maintenance would be a nightmare. Before that, I build a simulation tool, and again, nothing but C would do the job well for the core algorithms. Glue-code should of course be something else, but most certainly not Java (I uses Python, because embedding C in it works really well).
The bottom line is, and has been for a long time, that actual professionals know a lot of tools and use the best one for the job, while amateurs and low-level technicians (and web-coders usually fall in that class) use the hype-du-jour. And professionals find that there is a set of tools that never become obsolete. C is among them, C++ likely as well (although the language design is significantly worse than C) and Python, Perl, Ruby will also stick around. So will Haskell, come to think of it. Now the thing is, none of these is are languages that you can competently use unless you have significant experience and talent and relevant education. The software manufacturing segment that just want to make everything even cheaper will not employ these tools, because people that are good at using them are rare and expensive. Long-term, they are worth the money spend several times over, but (and that is the second problem) with the MBA-cretinization of management, nobody but really good managers take the long-term view anymore.
The people making these statements are all in the low-cost, low-skill segment. That is good for making a few people rich in straw-fires, but sustainable software development looks differently. They just likely have never seen it. I do agree that technician-level coders (i.e. solve known problems with known methods in known ways) will not have good job prospects longterm (which makes all these "learn to code" initiatives so braindead), but as soon as you are an actual engineer, that is not going to apply.
The posting said "counties". Unless you find a second one, it is still bullshit. And even if you do, it is still lying by misdirection, because LA is in no way a typical or representative "county". Shame on you!
That one is tricky. There are situations where you have to do it yourself to get the required trust-level. Of course, that does _not_ include designing the actual crypto algorithm, only a few people on this planet can do that at this time. But from your own example, you can see that even "what the NSA does" may not be good enough. (I know a few people that used to do work for the NSA. These people are only cooking with water, same as everybody else. Of course, they have a lot of water at their disposal, but they are human and they make mistakes. Also, the NSA is not an attractive employer for the best and brightest, they only get the second liege players.) But of course, this on very advanced, and students must know that they will need decades of experience if they ever be able to do this at all. That is why so much time is spent on attacks in a good security lecture: The students must get an intuition what can fail and what they are up against. Same as any good engineering lecture.
LED bulbs have high-voltage power electronics in there. By quality (e.g. Philips, but avoid Osram, they are still figuring it out) and they will live as long as advertised and be much, much cheaper overall than all alternatives. Buy cheap ones and you will get bulbs designed by people that do not have sufficient experience and will fail much sooner. LED bulbs do typically not fail in the LEDs, these are well-understood. They fail in the power-converters. This is not a surprise either.
That is not going to happen. Every sale lost due to lack of offering is money not made. At the moment, the content mafia still succeeds in hypnotizing everybody into thinking they are important, but the reality is that anybody providing entertainment is a beggar and dependent on the goodwill of their potential benefactors (i.e. customers).
Switzerland has a GDP of around $650B. There are only 6 _states_ in the US that have more and only 3 states have more than double that. Now, I did not fund a list of US GDP by country, but unless, for example, California manages to have about 25% of its total GDP in only one county, your numbers are complete bullshit.
Indeed. And if somebody is really not a team-player, because they do not have the personal maturity for it (Blocher), then that person will be removed again. The others often serve for decades and if they make a mess, chances are they will have to clean it up themselves. As a result, when the Swiss have votes every three months on things, including almost all laws, there is actual information being disseminated and the government has one recommendation that all 7 Bundesraete support, despite being comprised of several different parties. And they actually give rational arguments for that recommendation (in most cases).
Of course, even this can be subverted. For example, the Swiss recently voted themselves the beginnings of a surveillance-state without any good rational reason, just the usual fear-driven nonsense. But they can get rid of that again even against the government if they so chose, and this subversion is much harder to achieve than in other systems. And quite often those in power do not get what they want, especially if it costs a lot of money with no real gain.
I could not agree more. If you, for example, vote for a party that promised to do a real war and then they do it as result of being elected, you very much share the responsibility for the killing that ensues, just the same as if you had done it personally.
As to the matter at hand, a "war" is were people a lot of people get killed without actually deserving it individually. A war is not something you can overlook or misunderstand. If you are in a war, you _will_ know. That is another grown-up thing: Do not use very serious terms like "war", "terrorism", etc. for things that are something else.
It is also a pretty neat implementation error. I am going to use this in my security lecture this year as example on how to mess up key handling and why to not trust the OS API (unless you are sure it it good).
Indeed. What the NSA did here would be called treason in any non-government organization, because what they did massively helped enemies. They need to massively reduce the number of exploits they keep secret (I can understand that they want a few), they need to make very sure the exploits and exploit-code does not ever get stolen and they need to make sure the exploits they keep secret are both hard to find and hard to exploit. Unless and until they do that, they will indeed need to be considered an enemy of basically everybody.
Inevitably, how this dictates the type of people that get into computing is it currently attracts people comfortable with high levels of uncertainty, something engineers are not. My brother is as Nuclear Physicist, he likes to joke that for him, point A and point B are enough to define a straight line, but an engineer needs more data.
That captures it well really. I mostly work as an engineer (I am also a scientist, but part-time only), and when I do engineering, I not only want these two points, I want two more in addition so I have generous redundancy and _still_ have redundancy left even if one of the point fails. When doing Science, I am perfectly fine with using only two points;-)
These days, the US has just two things left: Being large and being very stupid. Hence electing Trump as president is fine, because he is an exceptionally appropriate representative of the US population.
Of course the NSA has the largest share of the blame, because they lost ready-to use 0-day exploit code. That is about the worst thing possible. The NSA is also to blame because they did not report the 0-day after a reasonable time, say 1 year or so.
That makes to major screw-ups or seriously criminal acts on the side of the NSA. MS puts out shoddy software, but a) everybody knows that and b) a lot of others do it to. So some, but not a lot of blame to MS. The the fuckups that used this code also have some blame, but they are just small-time criminals that would never be able to pull this off by themselves.
The NSA has almost all the blame here. And what they enabled is about comparable to a major terrorist attack in damage to critical infrastructure.
People like to ignore what they cannot fix, instead to learn how to fix it. And then they try to convince others that ignoring it is the right strategy in order to get emotional confirmation. I guess quite a few tribes and larger groups of people have failed due to that in history. Of course, this approach is anathema to any good engineer, because if we screw up, things break, sometimes spectacularly. Unfortunately, IT is not a proper engineering field today and many people working in it do not qualify as engineers from their mind-set and skills. They even argue that "programming" is not an engineering activity (it is using technology to build technological artifacts, so how can not be engineering?) and that leads to a lot of really bad coders, designers and architects.
Note that formal qualification is just one way to get that mind-set, and not an absolutely reliable one. I also support giving engineering degrees to people that can prove good skill and understanding what it means to be an engineer in their field, with some additional qualification required if needed. But also note that engineers that screw up badly and kill people or destroy a lot of value become personally liable. For example, there was a case recently in Germany, where a building engineer screwed up the calculation for the roof of an ice-skating hall (I think) some 15 years back. It did collapse under a snow-load in 2012 (I think) and killed somebody. That engineer was found guilty of involuntary man-slaughter because he made a calculation he was not qualified to do, screwed it up and did not ask anybody else to validate it and hence caused the roof to be defective. That is what it means to be an engineer: People can depend on the quality of your work, sometimes with their lives. When we have reached that state in software-engineering (and we will have to reach it, or everything will go to hell), then we will be a mature engineering discipline.
Malware authors steal from each other all the time. Sometimes you see a patchwork of different styles and skill-levels and constructs that make not any sense, except if a later attacker did not really understand the code he was modifying. Still interesting though.
The sub-set of Postscript used in PDF has Turing-power. All it needs is permissions and you can do with it whatever you want. Displaying stuff is just what it has default-permissions to do. This means you do only need a privilege escalation, and not the code execution vulnerability malware in non-executable formats needs.
An OS family is defined by its API (at least by sane people). So Linux is part of the UNIX family, and so is, for example, QNX. You do not have to like that, but you do not get to decide what reality is. I am well aware that Linux is not a UNIX as it does not share kernel-code with either SYS V or BSD. It is however UNIX-like and that makes it part of the family.
Fundamentalists like you harm everybody, including yourself. Please go away.
Slashdot Mirror
Indeed. It may be true that coding jobs on low skill level are often Java these days, but C is not vanishing at all. For example, I recently built a custom security filter component for a customer and while in theory it is possible to do this in Java (I think), it would be a lot of additional effort, the component would not perform well and maintenance would be a nightmare. Before that, I build a simulation tool, and again, nothing but C would do the job well for the core algorithms. Glue-code should of course be something else, but most certainly not Java (I uses Python, because embedding C in it works really well).
The bottom line is, and has been for a long time, that actual professionals know a lot of tools and use the best one for the job, while amateurs and low-level technicians (and web-coders usually fall in that class) use the hype-du-jour. And professionals find that there is a set of tools that never become obsolete. C is among them, C++ likely as well (although the language design is significantly worse than C) and Python, Perl, Ruby will also stick around. So will Haskell, come to think of it. Now the thing is, none of these is are languages that you can competently use unless you have significant experience and talent and relevant education. The software manufacturing segment that just want to make everything even cheaper will not employ these tools, because people that are good at using them are rare and expensive. Long-term, they are worth the money spend several times over, but (and that is the second problem) with the MBA-cretinization of management, nobody but really good managers take the long-term view anymore.
The people making these statements are all in the low-cost, low-skill segment. That is good for making a few people rich in straw-fires, but sustainable software development looks differently. They just likely have never seen it. I do agree that technician-level coders (i.e. solve known problems with known methods in known ways) will not have good job prospects longterm (which makes all these "learn to code" initiatives so braindead), but as soon as you are an actual engineer, that is not going to apply.
It is. I completely agree. For a country with so much potential to have so many poor people is disgraceful.
The posting said "counties". Unless you find a second one, it is still bullshit. And even if you do, it is still lying by misdirection, because LA is in no way a typical or representative "county". Shame on you!
That one is tricky. There are situations where you have to do it yourself to get the required trust-level. Of course, that does _not_ include designing the actual crypto algorithm, only a few people on this planet can do that at this time. But from your own example, you can see that even "what the NSA does" may not be good enough. (I know a few people that used to do work for the NSA. These people are only cooking with water, same as everybody else. Of course, they have a lot of water at their disposal, but they are human and they make mistakes. Also, the NSA is not an attractive employer for the best and brightest, they only get the second liege players.) But of course, this on very advanced, and students must know that they will need decades of experience if they ever be able to do this at all. That is why so much time is spent on attacks in a good security lecture: The students must get an intuition what can fail and what they are up against. Same as any good engineering lecture.
LED bulbs have high-voltage power electronics in there. By quality (e.g. Philips, but avoid Osram, they are still figuring it out) and they will live as long as advertised and be much, much cheaper overall than all alternatives. Buy cheap ones and you will get bulbs designed by people that do not have sufficient experience and will fail much sooner. LED bulbs do typically not fail in the LEDs, these are well-understood. They fail in the power-converters. This is not a surprise either.
That is not going to happen. Every sale lost due to lack of offering is money not made. At the moment, the content mafia still succeeds in hypnotizing everybody into thinking they are important, but the reality is that anybody providing entertainment is a beggar and dependent on the goodwill of their potential benefactors (i.e. customers).
Switzerland has a GDP of around $650B. There are only 6 _states_ in the US that have more and only 3 states have more than double that. Now, I did not fund a list of US GDP by country, but unless, for example, California manages to have about 25% of its total GDP in only one county, your numbers are complete bullshit.
Indeed. And if somebody is really not a team-player, because they do not have the personal maturity for it (Blocher), then that person will be removed again. The others often serve for decades and if they make a mess, chances are they will have to clean it up themselves. As a result, when the Swiss have votes every three months on things, including almost all laws, there is actual information being disseminated and the government has one recommendation that all 7 Bundesraete support, despite being comprised of several different parties. And they actually give rational arguments for that recommendation (in most cases).
Of course, even this can be subverted. For example, the Swiss recently voted themselves the beginnings of a surveillance-state without any good rational reason, just the usual fear-driven nonsense. But they can get rid of that again even against the government if they so chose, and this subversion is much harder to achieve than in other systems. And quite often those in power do not get what they want, especially if it costs a lot of money with no real gain.
I could not agree more. If you, for example, vote for a party that promised to do a real war and then they do it as result of being elected, you very much share the responsibility for the killing that ensues, just the same as if you had done it personally.
As to the matter at hand, a "war" is were people a lot of people get killed without actually deserving it individually. A war is not something you can overlook or misunderstand. If you are in a war, you _will_ know. That is another grown-up thing: Do not use very serious terms like "war", "terrorism", etc. for things that are something else.
It is also a pretty neat implementation error. I am going to use this in my security lecture this year as example on how to mess up key handling and why to not trust the OS API (unless you are sure it it good).
Indeed. What the NSA did here would be called treason in any non-government organization, because what they did massively helped enemies. They need to massively reduce the number of exploits they keep secret (I can understand that they want a few), they need to make very sure the exploits and exploit-code does not ever get stolen and they need to make sure the exploits they keep secret are both hard to find and hard to exploit. Unless and until they do that, they will indeed need to be considered an enemy of basically everybody.
Spoken like a true cave-man with zero understanding of the issue at hand.
Inevitably, how this dictates the type of people that get into computing is it currently attracts people comfortable with high levels of uncertainty, something engineers are not. My brother is as Nuclear Physicist, he likes to joke that for him, point A and point B are enough to define a straight line, but an engineer needs more data.
That captures it well really. I mostly work as an engineer (I am also a scientist, but part-time only), and when I do engineering, I not only want these two points, I want two more in addition so I have generous redundancy and _still_ have redundancy left even if one of the point fails. When doing Science, I am perfectly fine with using only two points ;-)
These days, the US has just two things left: Being large and being very stupid. Hence electing Trump as president is fine, because he is an exceptionally appropriate representative of the US population.
Of course the NSA has the largest share of the blame, because they lost ready-to use 0-day exploit code. That is about the worst thing possible.
The NSA is also to blame because they did not report the 0-day after a reasonable time, say 1 year or so.
That makes to major screw-ups or seriously criminal acts on the side of the NSA.
MS puts out shoddy software, but a) everybody knows that and b) a lot of others do it to. So some, but not a lot of blame to MS.
The the fuckups that used this code also have some blame, but they are just small-time criminals that would never be able to pull this off by themselves.
The NSA has almost all the blame here. And what they enabled is about comparable to a major terrorist attack in damage to critical infrastructure.
Just got it, and finally things work well ;-)
That works with Vivaldi? Excellent, thank you!
People like to ignore what they cannot fix, instead to learn how to fix it. And then they try to convince others that ignoring it is the right strategy in order to get emotional confirmation. I guess quite a few tribes and larger groups of people have failed due to that in history. Of course, this approach is anathema to any good engineer, because if we screw up, things break, sometimes spectacularly. Unfortunately, IT is not a proper engineering field today and many people working in it do not qualify as engineers from their mind-set and skills. They even argue that "programming" is not an engineering activity (it is using technology to build technological artifacts, so how can not be engineering?) and that leads to a lot of really bad coders, designers and architects.
Note that formal qualification is just one way to get that mind-set, and not an absolutely reliable one. I also support giving engineering degrees to people that can prove good skill and understanding what it means to be an engineer in their field, with some additional qualification required if needed. But also note that engineers that screw up badly and kill people or destroy a lot of value become personally liable. For example, there was a case recently in Germany, where a building engineer screwed up the calculation for the roof of an ice-skating hall (I think) some 15 years back. It did collapse under a snow-load in 2012 (I think) and killed somebody. That engineer was found guilty of involuntary man-slaughter because he made a calculation he was not qualified to do, screwed it up and did not ask anybody else to validate it and hence caused the roof to be defective. That is what it means to be an engineer: People can depend on the quality of your work, sometimes with their lives. When we have reached that state in software-engineering (and we will have to reach it, or everything will go to hell), then we will be a mature engineering discipline.
Malware authors steal from each other all the time. Sometimes you see a patchwork of different styles and skill-levels and constructs that make not any sense, except if a later attacker did not really understand the code he was modifying. Still interesting though.
The bookmark-management in Vivaldi still sucks (is it so hard to implement what Opera 12.x had?), but otherwise it is a fine browser.
Indeed. Turns out that "saving" money on IT security can become pretty expensive.
There are defenses that work. Just not on Windows. As usual, MS is far behind.
The sub-set of Postscript used in PDF has Turing-power. All it needs is permissions and you can do with it whatever you want. Displaying stuff is just what it has default-permissions to do. This means you do only need a privilege escalation, and not the code execution vulnerability malware in non-executable formats needs.
Indeed. And you cannot fix even more stupid, in particular the people who paid.
An OS family is defined by its API (at least by sane people). So Linux is part of the UNIX family, and so is, for example, QNX. You do not have to like that, but you do not get to decide what reality is. I am well aware that Linux is not a UNIX as it does not share kernel-code with either SYS V or BSD. It is however UNIX-like and that makes it part of the family.
Fundamentalists like you harm everybody, including yourself. Please go away.